Soundtracker origins, part 2: Welcome to Turrican, aah hahahaha

Temps de lecture / Reading time : 18 minutes.

It’s high time I write part two of this series of articles on the origins of Soundtracker, since the content itself has been lying in my inbox for well over two years now…

As a reminder: I’ve been writing about my « quest » of looking for the missing link between what seems to be the first « tracker-like » interface¹ and Karsten Obarski’s Ultimate Soundtracker tool, which introduced a cheap tracker interface² to Amiga musicians back in 1987.

So, where were we?

In part 1 of this series, we learnt more about Karsten Obarski, who became the « Father of the Soundtracker » at age 22. Through existing interviews, we got to understand where he came from, how he came to create his Ultimate Soundtracker tool on Amiga in 1987, why he called it quits a few months afterwards… and where he probably took his inspiration for The Ultimate Soundtracker.

Said inspiration was, by all accounts, an earlier tool named Soundmonitor, which German developer & musician Chris Hülsbeck wrote and released on Commodore 64 in 1986 — a year before Obarski’s own Ultimate Soundtracker. Hülsbeck was 18.

Soundmonitor V1.0 — Soundmonitor 1.0, released in October 1986.
I guess kids those days didn’t really need a manual.

Chris Hülsbeck went on to become world-famous by creating game music, not the least being the Turrican series of games³. He nowadays creates royalty-free music, and oversees orchestral renditions of the Turrican soundtrack, amongst other things. Looking at his Bandcamp page, you could say he keeps himself busy. Buy the vinyls!

Now you know why I chose that title for this article.

So that’s the status of our quest: Soundmonitor seems to have been the original tracker.

Or was it?

Writing computer game music in the 80’s

There’s much to learn about Chris Hülsbeck’s context at the time when he wrote Soundmonitor, back in 1986.

The context is: there were no music tools for the general public — even for seasoned amateurs. The first tools were very expensive and inaccessible to most musicians — let alone people who wrote computer game music.
The Fairlight CMI that I mentioned above⁴ was only used by a handful of people, for instance, and they were wealthy household names already: Peter Gabriel, Herbie Hancock⁵, Stevie Wonder, Kate Bush⁶, etc. Anecdote: The Miami Vice theme was composed on that tool ⁷.

Before the arrival of affordable tracking software such as Soundmonitor and Ultimate Soundtracker, game musicians had to rely on their programming skills to write their music. Game musicians were, for what it’s worth, programmers before anything else.
Eventually they did write their own music software, however crude and just for themselves, compiling their usual tricks into something easier to use on a regular basis, so as to be more productive.
But in the early days, they simply wrote their music in machine code, adding one hexadecimal value after the other in the code, in order to change volume, pitch⁸, or type of soundwave.

A successful Rob Hubbard.
I would totally trust this guy with my register. *Wink wink*.

Let’s take Rob Hubbard⁹, for instance. A professional studio musician by trade, he got interested in computers in the early 80’s, at roughly 27. He learned to program in Assembly language¹⁰, wrote music tooling, got hired as a game musician, and within a handful of years became of legend¹¹ of what is now known as « chiptune » music — music that exploits the sound chip included in computers at the time. Rob was a master at that.

Here’s an example of what great C64 game music sounded like in 1985:

You might want to lower your expectations of what « sound » means before clicking, just in case.

He became a master because he knew about analogue synthesizers before he got into computer music, and thus he understood the possibilities offered by their sound chip — namely, the equally legendary SID chip, whose sound is still very much appreciated today¹². Hubbard knew he could program the chip’s registers, and he did that aplenty.

Says the man himself in this video interview from 2017:

« I knew all the chromatic pitches pretty much all by heart in hexadecimal. 30x would be a C, 3Cx would be the octave above that, 48x would be the octave above that…
I knew all the numbers: I could get a machine dump and recognize exactly what was going on. »

« I used to know the SID chip inside & out. The filters were always different on the SID chip, you could never rely on them. »

« The Atari has an 8-bit register, so as you got higher in pitch, the resolution got less and less, it becomes very difficult to get certain notes in tune. You write your music around the fact that as you went higher, you could only rely on 3 or 4 pitches. »

« Three channels is basically all you had at your hand, so there was no choice. Later on I did manage to squeeze more out of it because I developed a digital channel as well. People were doing digital audio, using digital samples, and I was the first person who incorporated that into music, so that I could try to get a rock guitar in there with the SID chip, which was just unbelievable pain in the ass, because you’re using four bit, so the volume register (…)

Sounds like a tedious way of writing music — but not that far off from the Soundmonitor screen you see at the top of this article, ain’t it? And Soundmonitor was easy in comparison. Rob Hubbard, and the other genius composers of that golden era, wrote their music right in the code, in hexadecimal if need be.

His C64-Wiki page even says so:

He admits to having 3 ways of working:

write directly with the C64 by poking bytes using a machine code monitor;

write using a pen and paper;

sit at the keyboard and play until the ideas come out.

« Poking bytes using a machine code monitor ». Let that sink in. Now tracker programs look more visually pleasing, for sure.

And, well, the man himself described the musical landscape of the times in this interview:

Kenz: How did you go about composing your C64 tunes? Did you have a music editor that you used, as there were often rumours you used one you made yourself?
Rob: No, I just used an assembler and edited the source code – most people in those days did the same. There really wasn’t time to sit and write an editor, as there was so much work to do.
Commodore Zone interview.
We can also read in that page that he used Mikro Assembler by Andrew Trott, whose current homepage… mentions his interest for the Fairlight CMI. It’s all tied up, I tell ya!

And in another one:

Had you ever considered a music utility yourself?

« No, because I can’t think of any way to make my methods accessible enough to the average punter to make it worth while. »
Another Commodore Zone interview.

Music-making was thus confined to the programming elite.
Says Rob:

« It’s taken me [a] long while to develop my routines, and I’m not about to give them away! » (Rob hinted that one company which had made free with one of his demo disks might shortly regret having lifted his routines!).
Still that Commodore Zone interview.

Trackers were a few years away, because Chris Hülsbeck had yet to release his paradigm-shifting tool.

Let’s get to that.

The birth of Soundmonitor

Chris Hülsbeck was 18 when he « released »¹³ his Soundmonitor tool, in 1986.
What does a teenage geek do in his spare time? Painstakingly type down type-ins from computer magazines of course!

« Type-ins« ? Old farts like me remember buying magazines full of source code, hundreds of pages of them, that you would bring home and then type for hours on your computer, eventually compiling this code into a program or a small game — or a set of system-crashing errors if you typed something wrong. When it did work, that feeling of pride was immense: You had typed it; it was your work of art.
To many, this was the first introduction to programming — as well as English.

If you were lucky, the type-in used regular words¹⁴, making it easier to transcribe.
This one is from the French Amstrad Magazine n°01, from July 1985.

Of course the Internet existed since 1983 for some happy fews, and BBSs (bulletin board systems) were already a thing, but modems were still expensive in those days, and consumer-centric machines such as the C64, Atari ST or Amiga 500 weren’t equipped for global communication out of the box.

Magazines started to feature covermounts¹⁵ (a cover cassettes or floppy, eventually a cover CD-ROM) in the mid-80’s. But still, type-ins were quite popular in the 8-bit era.¹⁶

Alright alright alright, why am I rambling on about type-ins, you ask? Because that’s how Chris Hülsbeck got his start in the computer music world.

See, in early 1986, German computer magazine « 64’er » launched a music competition. You guessed it, the winner was 18-year-old Christopher Hülsbeck, with this piece of music:

Don’t click unless you are ready to face History, my friends.

If that doesn’t sound like much to your 2022 ears, read what the jury had to say:

It wasn’t easy for us to decide which song was the best – until we heard « Shades » by Chris Hülsbeck. A short breathless moment of silence, a restart, another short listen and then it was clear to us: this is the winner!

The truly unique composition of « Shades » deserves a lot of credit. Comparisons with the creations of professionals like Jean-Michel Jarre, Eberhard Schöner and similar synthesizer jugglers are not even that far-fetched.
« Shades » also compares favourably with the creations of the well-known C64 music professionals Rob Hubard (he wrote the music for the game « Thing on a Spring », for example) and Martin Galway (« Comic Bakery »).

About the future of the young Chris Hülsbeck one can assume: He can become one of them.

(…)
One more point that needs to be said: When we presented « Shades » for the first time to the C64 fans at the CeBIT fair in Hanover, many thought that this piece of music had been « borrowed » from some professional game. But this is not the case.
64’er, June 1986, translated.
If you’re nostalgic for shoulder pads, check out this 1986 CeBIT news report 🙂

The interesting part is that the whole piece of music was made available to the 64’er readers, through a 3-pages, 8-columnes long type-in fully in hexadecimal¹⁷.

Ah! That Amstrad type-in above doesn’t look so daunting after all, does it?

Even more interesting to us, right in the article presenting the winner, the editorial staff wrote: « Chris Hülsbeck, der Programmierer des Musikstückes »Shades« arbeitet derzeit an einem Editorprogramm » (« Chris Hülsbeck, the programmer of the music piece « Shades » is currently working on an editor program. »)

And, what do you know, five months later came this:

« Music… like never before »
64’er, October 1986. By the way, thanks a lot to The Internet Archive for hosting scans!

Behold! A vintage Chris Hülsbeck, seen here in his natural habitat.

When we started to evaluate our music competition a few months ago, the piece of music « Shades » by Chris Hülsbeck amazed us with some fantastic sounds. Until then, we were only used to something like this from professionals such as Rob Hubbard.

On an enclosed note, Chris Hülsbeck asked if we would like to have the music routine for publication. At that time Chris was programming the routine in an uncomfortable way with a machine language monitor. On the phone, he promised to write an editor around the « music master » (as Chris christened the music routine).

When the sound monitor was finished, we were so impressed that we wanted to make it available to other readers. It became the listing of the month.
64’er, October 1986, translated.

Follows a manual written by Hülsbeck himself, and then this, the type-in for the Soundmonitor program:

Fortunately, you could mail-order a cassette with the program on it. Phew. Enough type-ins already!

Thus was first released Soundmonitor, true father of the trackers. Or is it?

While I started my exploration thinking that Karsten Obarski came up with the tracker layout, it seems Ultimate Soundtracker mostly re-used the layout introduced by Soundmonitor, improving it thanks to the Amiga graphic abilities — and benefiting from the Paula chip, meaning 4 audio channels and the ability to use samples rather than synth sounds.

Now, to find out whether Chris Hülsbeck, in turn, found inspiration elsewhere…

More context setting¹⁸

Chris’ « manual », published in the 64’er magazine, offers some information of how Soundmonitor came to be, and how advanced it was at the time.

The sound chip of the C64 offers considerable possibilities, but unfortunately the comfort of programming leaves a lot to be desired.
There are already several programmes that support the programming of the SID. Most of them are designed in such a way that you can place notes on the corresponding staves or play sounds via the keyboard.
Even complete music studios are simulated, but all known programmes have a decisive disadvantage: the composed music can only be played if the complete programme is in the memory. If you want to place a piece of music in a self-written basic game, for example, you are ill-served with these programs.

That’s why a completely independent playback routine was programmed, which is called « Musicmaster ». With the appropriate data, results are achieved that can even surpass background music from professional games.
However, it would be extremely uncomfortable if the music data had to be entered with a normal machine language monitor, such as SMON. The piece of music « Shades » was composed in this way, which was time-consuming work.

For this reason, a « monitor » had to be developed that specifically supports the input of music data: the « sound monitor ». The programme differs from other sound editors in some essential features.
The main part of the programme, the playback routine, runs completely independently in interrupt. This means that the song can be listened to at any time, even during editing. This is an excellent control possibility, you can immediately hear what you are typing.
In addition, the sound monitor contains a « realtime-record » (recording of music while playing on the keyboard).
64’er, October 1986, translated.

Still, nothing about a possible predecessor, or an inspiration, or anything. Could it be that our quest ends here? That a passionate teenager simply came up with the tracker-format idea by himself, out of thin hair?

How could we know for sure?

Hm.

Let’s ask Chris Hülsbeck.

Interviewing Chris Hülsbeck

Unlike Karsten Obarski, who vanished shortly after releasing his Ultimate Soundtracker, Chris Hülsbeck has remained very active online: he makes royalty-free music through his Patreon, has all his music available on Bandcamp, maintains an active Twitter account, has released orchestral version of his most-known work, etc. You could say he keeps himself busy.

Chris Hülsbeck today, working from his motorhome in the US.
Funnily enough, in this photo (taken from his website), he displays two tools on the big screen: ProTracker 2.3d, and his very own TFMX editor (through the WinUAE emulator).

As you could see at the bottom of the first part of this series of article, I contacted Chris Hülsbeck back in 2019 through his website, not really expecting an answer — and getting one within 5 hours! Imagine being able to talk directly¹⁹ to someone whose music filled quite a few hours of your teenage years!

I’ll skip the overly long intro and context-setting from my email, since, well, I’ve already written that into this article 😅

To the first question, then!

XB: Did you come up with the tracker layout for Soundmonitor, or did you get inspiration from the Fairlight CMI (or any other tool)?
Were you in contact with Karsten Obarski, or any other « music-programmer » of the time?

CH: Let me start a bit earlier – despite having had 2 years of piano lessons when I was 5 years old, I never got to properly learn or appreciate musical notation.
By the time I was starting to compose and « program » music on the C64 (around age 16), I had developed my own musical understanding and language, otherwise I would probably have tried to represent the notes in a graphical way just like other musical software at the time.

I didn’t know about the Fairlight sequencer page at the time, but the inspiration for the note representation came from an early C64 midi sequencer by Steinberg, which didn’t display notes as symbols, but as a list of single letter note name, an optional « sharp » symbol if needed and a number for the octave. This made the most sense to me for computer music and the rest came together just by needing a simple layout for the song data and the patterns.

I didn’t meet other music software programmers until years later.
Private e-mail interview from June 2019²⁰.

🚨Alert, alert, we have an inspiration! 🚨
Chris used a tool by Steinberg, the editor of Cubase, the world-famous DAW²¹, and inventor of the de facto standard for digital audio plugins, VST²².

Back in the 80’s, Steinberg wasn’t the software superstar²³ that it is today. It was founded in 1984 by Karl Steinberg and Manfred Rürup, who were musicians and studio engineers, and their passion for the latest musical gear led them to write audio software, starting with the C64.

Their first best-selling software was « Pro-16 » from 1986, a tool able to manager up to 16 tracks of MIDI instruments…

Looks like a stepping system on the right, if I’m not mistaken…

… but their very fist tool was indeed named « Midi Multitrack Sequencer », and released in 1984²⁴.

Are we seeing some kind of sequencing here? Oh right, « sequencer » is even in the name, duh.

So, any of these two tools could have been the inspiration for Soundmonitor, it seems, since they date from before 1986.
It’s hard to find any resemblance with the tracker layout of Ultimate Soundtracker or even Soundmonitor, but the inspiration might as well come from the way those earlier tools operated.

Let’s try and get some details from Chris.

XB: So I understand you came up with the tracker-like layout, inspired by « an early C64 midi sequencer by Steinberg ».
Do you remember the name of that sequencer? Was it Pro 16, Trackstar, or even an earlier tool?

CH: I’m actually not quite sure… maybe it was the « MIDI Multitrack Sequenzer ». I did also work with and loved the Pro-16, but I don’t remember if I did before or after the Soundmonitor.
Private e-mail interview from June 2019.

So, MIDI Multitrack Sequencer is the next step in our quest, it seems. And it’s not a small one, since Steinberg is a whole other kind of thing today.

It’s one thing to contact a music composer who is very much present online; it’s a whole other thing to contact the co-founder and CEO of a company whose later tool, Cubase, is used by a large portion of professional and amateur musicians today.

I’ll try my luck anyway. But that will be a topic for another post here — I don’t know if you noticed, but this article is quite long already 🙂

Continuing with the interview

Of course I had a few more questions I wanted to ask to the man whose music livened quite a few of my teenage gaming evenings, so here goes!

XB: Your first tune, Shades, was as far as I understand, written directly in machine code — all for a music competition for the 64er magazine, which you won.
What was your journey between writing your own music, and writing your own editor?
And which came first, the song or the editor? 🙂

CH: The creation of Shades was actually quite painful, because I had to edit the hexadecimal numbers representing the note data and everything else directly in the computer memory using a program called a « monitor ».

So my idea was to make a similar program, but tailored to sound data, hence the Soundmonitor was born. When I told the writer for the magazine who was responsible for the music contest about my plans he was very excited and suggested that I should submit it to the sister magazine Happy Computer for their « listing of the month » and it was accepted right away.

XB: In an interview, Rob Hubbard said that he composed his C64 tunes directly in assembler, « most people in those days did the same. There really wasn’t time to sit and write an editor, as there was so much work to do. »
Obviously you took the time to do it. Is it because you weren’t a professional at the time — and thus had more free time?

Who else was making a popular editor at the time?

CH: I was still in school and on Summer break when I created the Soundmonitor, so I wasn’t bogged down with professional music jobs yet (as Hubbard said).

I honestly don’t know of any other editors during that time.
The first scene music tracker I became aware of was actually just a hack of the Soundmonitor called Rockmonitor, which added one track of a rather crude sample playback (different from my own sample playback system, which I never released to the public).

I like the idea of professionals not having the time to take a step back and write a proper tool, and being beaten to the finish line by a teenager on Summer break 🙂

XB: By the way, on your SID Anthology vol 1, Shades is only the second track. Does it mean Planet of War, the first track, is your first officially released track?

CH: Planet of War was composed and programmed before Shades, with a much simpler player, but it was actually released after Shades because of a delay to find a publisher for the game. But for the album I felt it needed to be chronological in terms of when it was composed.

XB: So, just as for Shades, I understand that all musicians at the time were actually programmers with a gift for melody — which might explain why they were few and fondly remembered (Hubbard, Galway, Follin, etc.)
How much of composing at the time was a result of trial and error?

CH: It was very inspiring if I would find a cool new sound or trick when composing, but often I would develop my melodies just with a piano sound on my keyboard synth and then translate that melody into the machine.
In any case it was a very tedious and technical way of creating music and I think that is one reason only a handful of people developed the skills that truly set them apart from the rest.

XB: The published source code for Soundmonitor takes 5 magazine pages long (albeit, with 3 columns). I find that impressively small for a full musical program! Did you already introduce optimizations in there?

CH: The listing is not a source code, but an actual compressed binary… I never released the source code itself and it’s actually a horrible piece of mess because I was so young that I didn’t care about readability or that I would need to ever go back and understand it… 😉

XB: As I understand it, one of the great step forward was that Soundmonitor could save and load files, with a separate playroutine, saving both time and space for other musicians. Was that also a creation of yours, or did you get inspired from others?

CH: That was just a very cool idea I had to give people the ability to put music into their own games or demos, even if they had only very basic programming skills. It also made it very easy for people to share their music and recipients didn’t even need the editor to listen…

XB: In an interview in 1992, Karsten Obarski said that « Even today’s trackers work in that same way, and still use the tone-event data structures which I invented. (Which is a very simple one.) » Do you understand what he means by « tone-event data structures »? I have an idea, but maybe that talks to you more?

CH: No exact idea, but I assume it’s got something to do with how his sequencer triggers notes and modulations. The Soundtracker format was much more streamlined compared to the Soundmonitor and had real memory management, so you couldn’t easily crash it by making mistakes, but I feel that also made it somewhat less flexible. Because trackers are easier to learn and use, the format has understandably surpassed the Soundmonitor.

XB: I found out that different waves produced different sounds (who would have thought? 🙂 ). Sine, square, triangle, sawtooth… Did you make use of these? Did Soundmonitor solely relied on such sounds, or could you import other sounds? (I understand TFMX could load samples, for instance).

CH: I took advantage of all the wave forms and possibilities that the SID chip offered and added sample playback to my own developer version of the Soundmonitor. By the time I developed my next tool TFMX, samples had actually been on the way out because they did not work on a new revision of the SID (until a different way was developed much later) and they took too much memory and CPU power to be included in the ever more complex games that were created. So TFMX doesn’t actually include sample playback, but instead I invented a new way of manipulating the SID chip with every screen cycle through something I called sound macros and it could result in complex instruments that almost sounded like samples (particularly drum sounds). This system lived on with later music I created for the Amiga, game consoles like Super Nintendo and Sega Megadrive as well as on the Nintendo 64 with a new tool that I developed with Factor 5 called MusyX. This sound tool still featured the concept of the TFMX sound macros and Nintendo actually bought a license to use it for their third party developers.

So that’s it for now!

Karsten Obarski’s Soundtracker took inspiration from Chris Hülsbeck’s Soundmonitor, who in turn took inspiration from Karl Steinberg’s MIDI Multitrack Sequencer. Phew!

What’s next? Will I find the missing link between the Fairlight CMI hardware workstation and later software audio tools? It’s a mystery!

Let’s just say that I closed my emails exchange with Chris Hülsbeck with this small spoiler:

As an aside, I’ve been interviewing the creator of the Fairlight CMI, and I think I know the exact moment when the whole « pattern of samples » format made the jump from a physical drum machine to the Fairlight’s sequencer, and who’s behind it! It’s a small thing, but this is exciting 🙂

See you next time!